Elsevier

Applied Soft Computing

Volume 55, June 2017, Pages 31-43
Applied Soft Computing

Bi-histogram equalization using modified histogram bins

https://doi.org/10.1016/j.asoc.2017.01.053Get rights and content

Highlights

  • The proposed BHEMHB improves conventional histogram equalization.

  • Histogram segmentation enables mean brightness preservation.

  • Histogram modification reduces domination effect of high-frequency histogram bins.

  • BHEMHB is tested using standard and cervical cell images.

  • Statistical analyses reveal improvement in entropy, PSNR and AMBE measurements.

Abstract

The shifting of image mean brightness and the domination of high-frequency bins during histogram equalization (HE) often result in the deteriorating quality of enhanced images and a considerable amount of information loss. This study proposes a novel approach based on bi-histogram equalization to improve its abilities in preserving information entropy and mean brightness. The proposed technique, named Bi-histogram Equalization using Modified Histogram Bins (BHEMHB), segments the input histogram based on the median brightness of an image and alters the histogram bins before HE is applied. Histogram segmentation enables mean brightness preservation, whereas the modification of histogram bins restricts the enhancement rate, thus minimizing the domination effects of high-frequency histogram bins. Simulation results show that BHEMHB significantly outperforms its peers in preserving the details and mean brightness of an image. The output image is visually pleasant with a natural appearance.

Introduction

Image enhancement remains one of the major concerns in the field of digital image processing. It manipulates an original input image to yield an image with better quality and improved interpretability. As an important image processing technique, image enhancement is extensively used in different applications, such as face recognition [1], watermarking [2], [3], medical image processing [4], and many others [5], [6], [7], [8], [9], [10]. Among the various image enhancement techniques that have been proposed, the technique based on conventional histogram equalization (CHE) is one of the most popular because of its simplicity and effectiveness. The main idea of CHE is to remap the gray levels of the image on the basis of probability density function (PDF). It works by flattening and stretching the dynamic range of the histogram, resulting in an overall enhancement of image contrast [11]. Despite its popularity, CHE suffers from a well-known drawback: it alters the original brightness of the input image. CHE always emphasizes image regions with higher number of gray level occurrences. These regions are frequently over-enhanced. By contrast, regions comprising a relatively small number of pixels may be eliminated, resulting in the so-called washed-out appearance. Several details in the image disappear as the gray levels of the output image decreases. Contrast stretching by CHE is also confined in certain dominated regions. The excessive merging of gray levels of the image results in false contours, which generate undesired artifacts and unnatural enhancement in the image [12]. Saturation problem also occurs where certain local areas are too bright in the output image, thus degrading the outlook of the image and resulting in information loss [13].

Substantial research has been undertaken to address the aforementioned drawbacks. In this work, the problems of mean brightness shifting and domination of high-frequency bins suffered by CHE are addressed. The rest of this paper is organized as follows. Related works on image enhancement, specifically techniques based on histogram equalization (HE), are discussed in Section 2. The research motivation followed by the proposed technique, namely, Bi-histogram Equalization using Modified Histogram Bins (BHEMHB), is outlined in Section 3. The data samples and performance measurement used are discussed in Section 4. The simulation results and discussions are outline in Section 5 using both qualitative and quantitative analyses. Finally, the conclusion of the work is presented in Section 6.

Section snippets

Related works

Several image enhancement techniques based on HE are reviewed for a more comprehensive study. The earliest work to overcome the problem of mean brightness shifting was proposed by Kim [14]. The proposed Brightness Preserving Bi-Histogram Equalization (BBHE) divides the histogram of the input image into two sub-histograms according to mean brightness of the image. Experimental results show that BBHE can reduce the saturation effect and avoid unnatural enhancement and annoying artifacts while

Research motivation and methodology

In this section, the research motivation is presented and the contribution is highlighted. Details of the proposed BHEMHB are also discussed.

Data samples and performance evaluations

This section discusses the data samples used (i.e., standard images from database and microscopic images) in this study. The qualitative and quantitative analyses carried out to evaluate the performance of BHEMHB are also presented.

Results and discussions

In this section, simulation results of the proposed BHEMHB and the seven HE-based techniques are presented. The performance of BHEMHB in enhancing microscopic medical images is then evaluated using cervical cell images.

Conclusion

A novel bi-histogram equalization technique, namely, Bi-histogram Equalization using Modified Histogram Bins (BHEMHB), is proposed in this paper to improve the ability of histogram equalization (HE) in terms of detail and mean brightness preservation. The novelty of BHEMHB is its idea of integrating histogram segmentation with the modification of histogram bins. The proposed technique successfully overcomes the shortcomings of HE, especially in mean brightness and detail preservation. This

Acknowledgements

The authors express their sincere thanks to the anonymous reviewers for their significant contributions to the improvement of the final paper. This study was partially supported by National Cancer Council Malaysia (MAKNA), Malaysia, under the project entitled “Development of an Intelligent Screening System for Cervical Cancer,” and by the Ministry of Higher Education (MOHE), Malaysia under MyPhD Scholarship.

Tang Jing Rui received her B. Eng. degree in Mechatronic Engineering with First Class Honors from Universiti Sains Malaysia (USM), Malaysia in 2012. She is currently a Ph.D. candidate at the School of Electrical and Electronic Engineering, USM and is connected with the Imaging and Intelligent System Research Team (ISRT). Her research interests include digital image processing and intelligent diagnostic systems.

References (63)

  • N. Al-Najdawi et al.

    Mammogram image visual enhancement, mass segmentation and classification

    Appl. Soft Comput.

    (2015)
  • A.S. Abdul Ghani et al.

    Enhancement of low quality underwater image through integrated global and local contrast correction

    Appl. Soft Comput.

    (2015)
  • A.S. Abdul Ghani et al.

    Underwater image quality enhancement through integrated color model with Rayleigh distribution

    Appl. Soft Comput.

    (2015)
  • S. Jenifer et al.

    Contrast enhancement and brightness preserving of digital mammograms using fuzzy clipped contrast-limited adaptive histogram equalization algorithm

    Appl. Soft Comput.

    (2016)
  • A. Kaur et al.

    Contrast enhancement for cephalometric images using wavelet-based modified adaptive histogram equalization

    Appl. Soft Comput.

    (2017)
  • K. Liang et al.

    A new adaptive contrast enhancement algorithm for infrared images based on double plateaus histogram equalization

    Infrared Phys. Technol.

    (2012)
  • J.R. Tang et al.

    Adaptive image enhancement based on bi-histogram equalization with a clipping limit

    Comput. Electr. Eng.

    (2014)
  • J.D. Chang et al.

    LBP-based fragile watermarking scheme for image tamper detection and recovery

    2013 IEEE International Symposium on Next-Generation Electronics (ISNE)

    (2013)
  • C. Li et al.

    A novel method of image enhancement via multi-scale fuzzy membership

  • M.Z. Iqbal et al.

    Satellite image resolution enhancement using dual-tree complex wavelet transform and nonlocal means

    IEEE Geosci. Remote Sens. Lett.

    (2013)
  • M.M. Riaz et al.

    Fuzzy C-means and principal component analysis based GPR image enhancement

    2013 IEEE Radar Conference (RadarCon13)

    (2013)
  • W. Roller et al.

    Technology based training for radar image interpreters

    2013 6th International Conference on Recent Advances in Space Technologies (RAST)

    (2013)
  • C.C. Chang et al.

    Texture synthesis approach using cooperative features, in Computer Graphics, Imaging and Visualization (CGIV)

    2013 10th International Conference

    (2013)
  • R.B. Ahire et al.

    Overview of satellite image resolution enhancement techniques

    2013 Tenth International Conference on Wireless and Optical Communications Networks (WOCN)

    (2013)
  • Y. Wang et al.

    Image enhancement based on equal area dualistic sub-image histogram equalization method

    IEEE Trans. Consum. Electron.

    (1999)
  • M. Abdullah-Al-Wadud et al.

    A dynamic histogram equalization for image contrast enhancement

    IEEE Trans. Consum. Electron.

    (2007)
  • Y.T. Kim

    Contrast enhancement using brightness preserving bi-histogram equalization

    IEEE Trans. Consum. Electron.

    (1997)
  • C.H. Ooi et al.

    Bi-histogram equalization with a plateau limit for digital image enhancement

    IEEE Trans. Consum. Electron.

    (2009)
  • S.D. Chen et al.

    Minimum mean brightness brror bi-histogram equalization in contrast enhancement

    IEEE Trans. Consum. Electron.

    (2003)
  • G. Maragatham et al.

    Contrast enhancement by object based histogram equalization

    World Congress on Information and Communication Technologies

    (2011)
  • S.D. Chen et al.

    Contrast enhancement using recursive mean-separate histogram equalization for scalable brightness preservation

    IEEE Trans. Consum. Electron.

    (2003)
  • Cited by (84)

    • A deep journey into image enhancement: A survey of current and emerging trends

      2023, Information Fusion
      Citation Excerpt :

      It is crucial to keep the image's average brightness and naturalness when enhancing it. The bi-histogram equalisation utilising modified histogram bins (BHE-MHBs) [44] method aids in addressing of prior problems. The input histogram is splits to two sub-histograms using the median value which are then altered to avoid dominant effects of high intensity histogram bins.

    View all citing articles on Scopus

    Tang Jing Rui received her B. Eng. degree in Mechatronic Engineering with First Class Honors from Universiti Sains Malaysia (USM), Malaysia in 2012. She is currently a Ph.D. candidate at the School of Electrical and Electronic Engineering, USM and is connected with the Imaging and Intelligent System Research Team (ISRT). Her research interests include digital image processing and intelligent diagnostic systems.

    Nor Ashidi Mat Isa received his B. Eng. degree in Electrical and Electronic Engineering with First Class Honors and Ph.D. degree in Electronic Engineering (majoring in Image Processing and Artificial Neural Network) from USM in 1999 and 2003 respectively. Currently, he serves as a Professor and lectures at the School of Electrical and Electronic Engineering, USM. His research interests include image processing, neural network, intelligent systems, biomedical engineering, and algorithms.

    View full text